5. Discussion

Several research works, studies have been conducted for analyzing the environment impact of CAVs [16–18].

Collaborative research from Argonne National Laboratory, National Renewable Energy Laboratory (NREL), Oak Ridge National Laboratory, and U.S. Department of Energy illustrates the overall system VMT and fuel consumption effects due to vehicle automation [16].

In the study [16], the researchers have considered four scenarios: conventional vehicles, partial autonomy, full autonomy with ridesharing, and full autonomy without ridesharing, and associates upper (UB) and lower (LB) bounds with the latter two. Figure 3 shows the upper bound and lower bound estimates on total U.S. Light-duty vehicle (LDV) fuel consumption for various CAV scenarios compared with the base Conventional scenario.

In Figure 3, it can be seen that there is large variation in the results; the total U.S. LDV fuel consumption (billion gallons per year) ranges from approximately 64% decrease in case of full autonomy with ridesharing LB up to approximately 205% increase in case of full autonomy without ridesharing UB.

Viewing the deployment of the CAVs (partial or full autonomy), the effects of travel demand and fuel efficiency were further investigated.

Possible changes in travel demand due to connectivity and increased vehicle automation are uncertain. The potential effects of CAVs on travel demand are classified into the following categories: less hunting for parking; easier travel; increased travel by under-served populations; mode shift from walking, transit and regional air; increase in ridesharing; and Increased empty miles traveled by automated vehicles [16].

Similarly, connectivity and vehicle automation have the potential to impact driving patterns, vehicle design as well as fuel efficiency. However, the impacts are uncertain. The categories of potential energy impact may include: drive profile and traffic flow smoothing; lesser congestion; efficient V2I/I2V communication; collision avoidance; platooning; and vehicle/powertrain resizing [16].

The researchers have elaborated the methodology that accommodates the abovementioned travel demand and fuel consumption effect assumptions to estimate

Connected Autonomous Electric Vehicles as Enablers for Low-Carbon Future DOI: http://dx.doi.org/10.5772/intechopen.84287

#### Figure 3.

More efficient driving in CAEVs can be achieved through a variety of mechanisms, including optimal driving cycle, dynamic eco-routing, traffic flow smooth-

A wide-scale deployment of CAEVs could facilitate vehicle platooning that could lead to improved aerodynamics. These advances in CAEVs could lead to consider-

Eco-driving may include route planning, trajectory optimization, and driving behavior improvement. And it is an effective way to reduce vehicle fuel consump-

Platooning is based on Cooperative ACC (CACC) technologies that use V2V communication to enable constant time-gap following and ad hoc joining and leaving the platoon. Platooning dynamically chains CAEVs to maximize fuel efficiency. Platooning is appealing due to the fact that it provides energy savings from aerodynamic drafting, more stable vehicle following dynamics, reduced traffic flow

Several research works, studies have been conducted for analyzing the environ-

Collaborative research from Argonne National Laboratory, National Renewable Energy Laboratory (NREL), Oak Ridge National Laboratory, and U.S. Department of Energy illustrates the overall system VMT and fuel consumption effects due to

In the study [16], the researchers have considered four scenarios: conventional vehicles, partial autonomy, full autonomy with ridesharing, and full autonomy without ridesharing, and associates upper (UB) and lower (LB) bounds with the latter two. Figure 3 shows the upper bound and lower bound estimates on total U.S. Light-duty vehicle (LDV) fuel consumption for various CAV scenarios compared

In Figure 3, it can be seen that there is large variation in the results; the total U.S. LDV fuel consumption (billion gallons per year) ranges from approximately 64% decrease in case of full autonomy with ridesharing LB up to approximately 205%

Viewing the deployment of the CAVs (partial or full autonomy), the effects of

Possible changes in travel demand due to connectivity and increased vehicle automation are uncertain. The potential effects of CAVs on travel demand are classified into the following categories: less hunting for parking; easier travel; increased travel by under-served populations; mode shift from walking, transit and

Similarly, connectivity and vehicle automation have the potential to impact driving patterns, vehicle design as well as fuel efficiency. However, the impacts are uncertain. The categories of potential energy impact may include: drive profile and traffic flow smoothing; lesser congestion; efficient V2I/I2V communication; colli-

The researchers have elaborated the methodology that accommodates the abovementioned travel demand and fuel consumption effect assumptions to estimate

regional air; increase in ridesharing; and Increased empty miles traveled by

sion avoidance; platooning; and vehicle/powertrain resizing [16].

tion and achieve significant reduction in carbon emissions.

disturbances as well as potential safety improvements.

ing, and speed harmonization.

able improvements in fuel economy.

Research Trends and Challenges in Smart Grids

4.4 Eco-driving and platooning

5. Discussion

ment impact of CAVs [16–18].

with the base Conventional scenario.

increase in case of full autonomy without ridesharing UB.

travel demand and fuel efficiency were further investigated.

vehicle automation [16].

automated vehicles [16].

18

Total U.S. LDV fuel consumption for various CAV scenarios compared with the base conventional scenario [16].

national-level fuel use impacts. Table 1 shows the notations used in the equations for the computation of national fuel consumption impacts [16].

Referring to Table 1, the impacts r<sup>t</sup> i,j, q<sup>t</sup> i,j are the fractional changes in the fuel consumption per mile over and above the fuel consumption per mile including all impacts considered earlier, that is, [16]

$$r\_t^{i,j} = \left(\frac{F\mathbf{C}\_t^{i,j}}{F\mathbf{C}\_{t-1}^{i,j}}\right) - \mathbf{1},\\q\_t^{i,j} = \left(\frac{F\mathbf{C}\_t^{i,j}}{F\mathbf{C}\_{t-1}^{i,j}}\right) - \mathbf{1} \tag{1}$$

and analogously for p<sup>t</sup> i,j, and s<sup>t</sup> i,j:

$$p\_t^{i,j} = \left(\frac{\text{VMT}\_t^{i,j}}{\text{VMT}\_{t-1}^{i,j}}\right) - \mathbf{1}, s\_t^{i,j} = \left(\frac{\text{VMT}\_t^{i,j}}{\text{VMT}\_{t-1}^{i,j}}\right) - \mathbf{1} \tag{2}$$

Using the notations in Table 1, the baseline conventional fuel use in the U.S. (without CAVs) is calculated as: [16]

$$\sum\_{i \in I, j \in \mathcal{J}} \left( \mathbf{VMT}\_0^{i,j} \* \mathbf{FC}\_0^{i,j} \right) \tag{3}$$


Table 1.

Notations used in equations for the computation of National fuel consumption impacts [16].

Consequently, the total fuel consumptions under partial automation and full automation scenarios can be calculated as follows: [16]

$$\sum\_{i \in I, j \in J} \left( \left( \text{VMT}\_0^{i,j} \prod\_{t \in T} \left( \mathbf{1} + p\_t^{i,j} \right) \right) \* \left( \text{FC}\_0^{i,j} \prod\_{t \in T} \left( \mathbf{1} + r\_t^{i,j} \right) \right) \right) \tag{4}$$

$$\sum\_{i \in I, j \in J} \left( \left( \mathbf{VMT}\_0^{i,j} \prod\_{t \in T} \left( \mathbf{1} + s\_t^{i,j} \right) \right) \* \left( \mathbf{FC}\_0^{i,j} \prod\_{t \in T} \left( \mathbf{1} + q\_t^{i,j} \right) \right) \right) \tag{5}$$

Based on the EPA Motor Vehicle Emission Simulator (MOVES) model values for U.S. national averages, the fraction of VMT on city and highway roads at peak and non-peak hours are estimated.

The analysis considers that the average fuel economy of LDVs to be 26.9 miles per gallon. Particularly, the analysis uses the relationship between city/highway fuel economy values with the combined fuel economy for the computation of the average city and highway fuel economy.

Additionally, it is assumed that traffic congestion occurs during peak hours and free flow driving occurs during non-peak hours. The analysis accounts the adjustment factors that have been suggested to calculate differences in fuel consumption (or GHG emission) under congestion and free flow driving and applies those adjustment factors in order to compute fuel economy and fuel consumption values during peak and non-peak hours. Table 2 shows the assumptions of VMT in percent, fuel economy and fuel consumption for a baseline conventional vehicle under various road types and time of day [16].

• BASE-AEO is a scenario that is based on EIA's Annual Energy Outlook (AEO)

• BASE-ADOPT is a scenario that is based on AEO 2017 inputs with projected vehicle sales shares from NREL's Automotive Deployment Options Projection

• CACC-AEO is a scenario with Cooperative Adaptive Cruise Control (CACC)

• CACC-ADOPT is a scenario with CACC applied to the BASE-ADOPT case;

• AutoTaxi-AEO is a scenario with automated taxis penetration projections

• AutoTaxi-ADOPT is a scenario with automated taxis applied to the BASE-

In Figure 5, it can be depicted that AutoTaxi scenerios (i.e., AutoTaxi-AEO, AutoTaxi-ADOPT) will have considerable energy impacts in compared with BASE use cases (i.e., BASE-AEO, BASE-ADOPT) in the future. For instance, by 2040, the US total LDV fuel consumption of CAVs decreases by 5% in case of AEO and 5.5%

The paper [18] is based the well-established ASIF framework, which expresses

<sup>∗</sup> F (6)

transport carbon emissions in terms of the major drivers. The formulation for

<sup>E</sup> <sup>¼</sup> <sup>A</sup><sup>∗</sup> <sup>S</sup><sup>∗</sup> <sup>I</sup>

Figure 5 shows U.S. total LDV fuel consumption for various scenarios (i.e., BASE-AEO, BASE-ADOPT, CACC-AEO, CACC-ADOPT, AutoTaxi-AEO,

AutoTaxi-ADOPT) for a certain time period from 2015 to 2050 [17].

carbon emissions (E) can be stated in the following equation: [18].

penetration projections applied to the BASE-AEO case;

Vehicle-level fuel consumption per mile impact under various vehicle automations [16].

Connected Autonomous Electric Vehicles as Enablers for Low-Carbon Future

DOI: http://dx.doi.org/10.5772/intechopen.84287

applied to the BASE-AEO case; and

2017 Reference case;

Tool (ADOPT);

Figure 4.

ADOPT case.

in case of ADOPT.

21

The preceding formulations (Eqs. (1)–(5)), along with the assumptions from Table 2, yield the fuel consumption per mile impacts of various vehicle automation technologies at a national level, as shown in Figure 4. The effects are distinguished by partial and full automation CAVs [16].

In Figure 4, it can be seen that the adoption of full automation CAVs may have significant productive energy impacts. Typically, the increased fuel savings due to various categories are as follows: for vehicle/powertrain resizing, fuel saving is 0%–50%; for drive profile and traffic flow smoothing, it is 6.5%–16%; for platooning, it is 3%–5%; for intersection V2I/I2V communication, it is 2%–4%; and for collision avoidance, it is 0.2%–2.2%.

In the paper [17], the researchers have presented a methodological approach for refining this wide range of estimated fuel consumption.

The researchers have utilized a framework that accounts for energy impacts at the vehicle level, projected adoption levels, and changes in VMT in order to estimate national level fuel consumption impacts of CAVs. And they have considered several scenarios [17]


Table 2.

VMT percent, fuel economy, fuel consumption assumed for conventional vehicle by road type and time of day [16].

Connected Autonomous Electric Vehicles as Enablers for Low-Carbon Future DOI: http://dx.doi.org/10.5772/intechopen.84287

#### Figure 4.

Consequently, the total fuel consumptions under partial automation and full

i,j t

� � � � � �

� � � � � �

Based on the EPA Motor Vehicle Emission Simulator (MOVES) model values for U.S. national averages, the fraction of VMT on city and highway roads at peak and

The analysis considers that the average fuel economy of LDVs to be 26.9 miles per gallon. Particularly, the analysis uses the relationship between city/highway fuel economy values with the combined fuel economy for the computation of the aver-

Additionally, it is assumed that traffic congestion occurs during peak hours and free flow driving occurs during non-peak hours. The analysis accounts the adjustment factors that have been suggested to calculate differences in fuel consumption (or GHG emission) under congestion and free flow driving and applies those adjustment factors in order to compute fuel economy and fuel consumption values during peak and non-peak hours. Table 2 shows the assumptions of VMT in percent, fuel economy and fuel consumption for a baseline conventional vehicle under

The preceding formulations (Eqs. (1)–(5)), along with the assumptions from Table 2, yield the fuel consumption per mile impacts of various vehicle automation technologies at a national level, as shown in Figure 4. The effects are distinguished

In Figure 4, it can be seen that the adoption of full automation CAVs may have significant productive energy impacts. Typically, the increased fuel savings due to various categories are as follows: for vehicle/powertrain resizing, fuel saving is 0%–50%; for drive profile and traffic flow smoothing, it is 6.5%–16%; for platooning, it is 3%–5%; for intersection V2I/I2V communication, it is 2%–4%;

In the paper [17], the researchers have presented a methodological approach for

The researchers have utilized a framework that accounts for energy impacts at the vehicle level, projected adoption levels, and changes in VMT in order to estimate national level fuel consumption impacts of CAVs. And they have considered

(U.S. MPG)

VMT percent, fuel economy, fuel consumption assumed for conventional vehicle by road type and time of

27 35 0.0286

Highway, peak hours 18 29.7 0.0337

City, peak hours 22 21.4 0.0467 City, non-peak hours 33 25.2 0.0397

∗ FCi,j 0 Y

∗ FCi,j 0 Y t∈T

<sup>t</sup>∈<sup>T</sup> 1 þ r

1 þ q i,j t

i,j t

Fuel consumption rate (U.S. GPM)

(4)

(5)

<sup>t</sup>∈<sup>T</sup> 1 þ p

1 þ s i,j t

� � � �

� � � �

automation scenarios can be calculated as follows: [16]

0 Y

0 Y t∈T

<sup>∑</sup><sup>i</sup>∈I,j∈<sup>J</sup> VMTi,j

Research Trends and Challenges in Smart Grids

<sup>∑</sup><sup>i</sup>∈I,j∈<sup>J</sup> VMTi,j

non-peak hours are estimated.

age city and highway fuel economy.

various road types and time of day [16].

by partial and full automation CAVs [16].

and for collision avoidance, it is 0.2%–2.2%.

several scenarios [17]

Highway, non-peak

hours

Table 2.

day [16].

20

refining this wide range of estimated fuel consumption.

Road type/time of day VMT % Fuel economy

Vehicle-level fuel consumption per mile impact under various vehicle automations [16].


Figure 5 shows U.S. total LDV fuel consumption for various scenarios (i.e., BASE-AEO, BASE-ADOPT, CACC-AEO, CACC-ADOPT, AutoTaxi-AEO, AutoTaxi-ADOPT) for a certain time period from 2015 to 2050 [17].

In Figure 5, it can be depicted that AutoTaxi scenerios (i.e., AutoTaxi-AEO, AutoTaxi-ADOPT) will have considerable energy impacts in compared with BASE use cases (i.e., BASE-AEO, BASE-ADOPT) in the future. For instance, by 2040, the US total LDV fuel consumption of CAVs decreases by 5% in case of AEO and 5.5% in case of ADOPT.

The paper [18] is based the well-established ASIF framework, which expresses transport carbon emissions in terms of the major drivers. The formulation for carbon emissions (E) can be stated in the following equation: [18].

$$\mathbf{E} = \mathbf{A}^\* \; \mathbf{S}^\* \; \mathbf{I}^\* \; \mathbf{F} \tag{6}$$

It can be seen that several mechanisms may yield substantial reduction in energy use and carbon emissions, while others may have negative impacts. For instance, utilization of eco-driving, platooning, congestion mitigation, de-emphasized vehicle performance, lower crash risk, vehicle right-sizing, car-sharing and on-demand mobility, and reduced infrastructure footprint of automated vehicles may contribute to the improved energy efficiency of AVs. However, the increase in VMT due to lower travel costs, new user groups (youth, elderly, disabled), higher highway speeds, and increased vehicle features may increase the carbon footprints of AVs. Table 4 provides abridged version of automation scenarios along with estimated ASIF multipliers for each effect. The scenarios vary in terms of levels of vehicle automation, effectiveness of the above-mentioned mechanisms in altering energy intensity, the degree of travel cost reductions, and the magnitude of travel

Connected Autonomous Electric Vehicles as Enablers for Low-Carbon Future

The authors have conducted scenario analysis that shows vehicle automation may reduce energy use and GHG emissions by half in best-case scenario, or double them in a worst-case scenario, depending on the effects that come to dominate. Consequently, the outcomes depend on which scenarios prevail and proactive pol-

Overall energy and environmental implications of CAEVs in future will depend

icy making is essential to steer the technology toward energy efficiency.

Estimated ranges of energy impacts of vehicle automation in respect of various mechanisms [18].

demand [18].

• Platooning.

Figure 6.

23

on following influencing mechanisms:

DOI: http://dx.doi.org/10.5772/intechopen.84287

• Changed mobility services.

• Travel-cost implications.

• Vehicle utilization.

• Vehicle operation (i.e., eco-driving).

• Energy-saving algorithms and vehicle design.

• Electrification using renewable energy resources.

### Figure 5.

U.S. total LDV fuel consumption scenarios for a certain time period from 2015 to 2050 [17].

where, A is activity level; S is modal share; I is energy intensity; and F is fuel carbon content.

The ASIF framework functions as a tool to organize various anticipated mechanisms through which vehicle automation may affect energy consumption and carbon emissions. Each driving factor on the right hand side of Eq. (6) can be considerably affected by the use of vehicular automation and thus fuel consumption and carbon emissions. Table 3 illustrates a concise version of presumed mechanisms for energy impacts of automated vehicles (refer to [18] for details).

Figure 6 illustrates estimated ranges of possible energy impacts of vehicle automation in respect of various mechanisms [18].


#### Table 3.

Presumed mechanisms for energy impacts of AVs [18].

Connected Autonomous Electric Vehicles as Enablers for Low-Carbon Future DOI: http://dx.doi.org/10.5772/intechopen.84287

It can be seen that several mechanisms may yield substantial reduction in energy use and carbon emissions, while others may have negative impacts. For instance, utilization of eco-driving, platooning, congestion mitigation, de-emphasized vehicle performance, lower crash risk, vehicle right-sizing, car-sharing and on-demand mobility, and reduced infrastructure footprint of automated vehicles may contribute to the improved energy efficiency of AVs. However, the increase in VMT due to lower travel costs, new user groups (youth, elderly, disabled), higher highway speeds, and increased vehicle features may increase the carbon footprints of AVs.

Table 4 provides abridged version of automation scenarios along with estimated ASIF multipliers for each effect. The scenarios vary in terms of levels of vehicle automation, effectiveness of the above-mentioned mechanisms in altering energy intensity, the degree of travel cost reductions, and the magnitude of travel demand [18].

The authors have conducted scenario analysis that shows vehicle automation may reduce energy use and GHG emissions by half in best-case scenario, or double them in a worst-case scenario, depending on the effects that come to dominate. Consequently, the outcomes depend on which scenarios prevail and proactive policy making is essential to steer the technology toward energy efficiency.

Overall energy and environmental implications of CAEVs in future will depend on following influencing mechanisms:


where, A is activity level; S is modal share; I is energy intensity; and F is fuel

considerably affected by the use of vehicular automation and thus fuel consumption and carbon emissions. Table 3 illustrates a concise version of presumed mechanisms for energy impacts of automated vehicles (refer to [18] for details).

Figure 6 illustrates estimated ranges of possible energy impacts of vehicle auto-

Congestion mitigation I ve 1–4 Moderate to high Eco-driving I ve 1–4 Any Platooning I ve 2–4 Any Higher highway speeds I +ve 1–4 Moderate to high De-emphasized performance I ve 3,4 Any Improved crash avoidance I ve 2–4 Very high Vehicle right-sizing I ve 3,4 High to very high Increased features I +ve 3,4 Any

Demand from New user groups A,S +ve 3,4 Any Changed mobility services A,S ve 3,4 Any Potential for low-carbon transition F ve 3,4 High

Direction of effect

Automation level

A,S +ve 1–4 Any

Penetration level

bon emissions. Each driving factor on the right hand side of Eq. (6) can be

element

U.S. total LDV fuel consumption scenarios for a certain time period from 2015 to 2050 [17].

mation in respect of various mechanisms [18].

Mechanism ASIF

Research Trends and Challenges in Smart Grids

Presumed mechanisms for energy impacts of AVs [18].

Demand due to travel cost

reduction

Table 3.

22

The ASIF framework functions as a tool to organize various anticipated mechanisms through which vehicle automation may affect energy consumption and car-

carbon content.

Figure 5.


Figure 6. Estimated ranges of energy impacts of vehicle automation in respect of various mechanisms [18].


reductions in GHG emissions from transportation and are at the forefront of this rapid transformation in transportation. The CAEVs have the great potential to operate with even higher vehicle efficiency, if they are charged using the electricity generated from renewable energy sources that will significantly reduce emissions as

Connected Autonomous Electric Vehicles as Enablers for Low-Carbon Future

transportation as well as improvement of vehicle efficiency.

The authors declare no conflict of interest.

This book chapter provides the energy synergy of combining vehicular automation, vehicle electrification, and vehicular connectivity along with appealing ondemand mobility services. It also furnishes understanding of the potential environmental implications of CAEV technologies. Using several studies, the chapter highlights the analysis of environmental impacts including GHG emissions due to the

This research work is supported by Smart Grid Fund (SGF), Ministry of Energy,

The Ontario Government and Canada Research Chair (CRC) Fund, Canada.

Some of the useful acronyms and abbreviations used in this book chapter:

well as dependency on fossil fuels.

DOI: http://dx.doi.org/10.5772/intechopen.84287

Acknowledgements

Conflict of interest

Acronyms and abbreviations

GHG greenhouse gas ZEV zero-emission vehicle AV autonomous vehicle

V2V vehicle to vehicle V2I vehicle to infrastructure VMT vehicle mile traveled MaaS Mobility-as-a-Service

25

ICE internal combustion engine

CAV connected and autonomous vehicle

CAEV connected and autonomous electric vehicle

#### Table 4.

Automation scenarios along with estimated ASIF multipliers for each effect (abridged version) [18].


Cybersecurity in CAEV networks is one of the active research areas [19–22]. Cybersecurity of CAEVs is essential for smart and sustainable development of a low-carbon city, since it may provide safety and social stability as well as economic sustainability.

Provisioning security and privacy in low-carbon smart mobility is crucial as without secure communications between CAEVs and remote systems may yield susceptibility to malicious attacks. For instance, compromised global position system (GPS) data affects the localization of CAEVs that may lead to traffic instability and/or hazardous accidents. Similarly, information shared among CAEVs in cooperative driving should be protected from any cyberattacks not only to guarantee road traffic safety but also to preserve privacy of CAEVs and other participating entities.

Current researches on CAEVs focus to identify cyber threats and vulnerabilities as well as to design strategies for preventing damages caused by these cyberattacks. Cyber threats and attacks studied include passive attacks such as eavesdropping, interception attack, traffic analysis and active attacks including impersonation attack, spoofing attack, replay attack, Sybil attack, jamming attack, message tampering [22]. Basically, requirements for cybersecurity solutions for CAEV networks may range from authentication, non-repudiation, integrity, to confidentiality.

Several open issues that should be addressed in future include [22]: in-vehicle security; security challenges in low-carbon smart cities; safety and security countermeasure consistency; and safe and secure mixed traffic systems.
